Mute ydb/core/kqp/ut/federated_query/datastreams/HasPathDatastreams.ExternalTableMatches in main

[ydbot]

Mute:
ydb/core/kqp/ut/federated_query/datastreams/HasPathDatastreams.ExternalTableMatches

Branch:
main

Build type:
relwithdebinfo

Owner: TEAM:@ydb-platform/fq

About this issue

• Auto-created by the mute workflow.
• Bot closes when all listed tests are unmuted or no longer exist.
• Close as Completed to trigger fast-unmute.
• mute_rules.md — full details.

Summary history:
HasPathDatastreams.ExternalTableMatches: Flaky(2026-07-14)->Passed(2026-07-15)->Flaky(2026-07-17) last 1 days, at 2026-07-14:2026-07-17: success_rate 96.6%, p-86, f-3, m-0, s-0, total-89

Test run history: link

More info in dashboard

[ydbot]

LLM instructions

Classify: TEST_ISSUE | YDB_ISSUE | TEST_INFRA_ISSUE
- YDB_ISSUE: product bug under ydb/core, ydb/library, etc.
- TEST_ISSUE: test logic, stale refs in helpers/workloads, missing waits. Fixture/harness PR
  that exposed a latent test bug → TEST_ISSUE; Responsible = fix path, not PR author.
- TEST_INFRA_ISSUE: shared harness/runner/CI broken for many unrelated tests. One directory /
  one test pattern → usually TEST_ISSUE, not TEST_INFRA_ISSUE.

Find root cause and introducing commit/PR; severity HIGH/MEDIUM/LOW; propose a fix.
In Root Cause list Suspect PR(s) most-likely-first (#12345 + why; tie to code/regression window).

Summary history (success_rate, p-*, f-*, total-*, dates): 50% often = runs before vs after a
merge, not a timing race. Compare failing tests with sibling tests (same fixture, still pass).

After fixture stop_driver(): "Driver was stopped" / "Topic client closed" → check stale driver
snapshot in workload/helpers before suggesting retries, sleeps, or graceful-shutdown revert.

VERIFY/SANITIZER — use status_description in exactly 3 lines:
  VERIFY failed (<ISO_UTC>): / ydb/.../file:line / assertion text (no stack frames, no +0x...)

Responsible (NOT issue-body Owner: — that is mute assignee only):
Pick team for the root-cause fix (first match):
  a) Fix path → .github/TESTOWNERS longest prefix (wins if PR exposed latent test bug)
  b) YDB_ISSUE → TESTOWNERS for broken ydb/... path
  c) Harness itself wrong → introducing PR author team
  d) Pure CI/Actions, no product commit → area/engineering
Map slug → area/... via .github/config/owner_area_mapping.json
Rules: root cause > Owner; commit from A exposing bug in B → B fixes; use area/queryprocessor not area/@ydb-platform/...

Reply exactly:
## Analysis
### Classification:
#### Error type: <TYPE>
#### Severity: <LEVEL>
#### Responsible: area/<team>
### Root Cause
<what broke>
Suspect PR(s), ordered by causality (most likely first):
1. #<pr> — <why>
### Suggested Fix
<fix or investigation>
### Summary
<2-4 sentences: failure, root cause, who fixes, next step>

Links to recent failed test runs

test	status	type	last_run	history	run	stderr	stdout	log	logsdir
HasPathDatastreams.ExternalTableMatches	failure	N/A	2026-07-17 14:11:02 UTC	link	link	link	link	n/a	link

[robot-vibe-db]

Analysis

Classification:

Error type: TEST_ISSUE

Severity: LOW

Responsible: area/fq

Root Cause

The test HasPathDatastreams.ExternalTableMatches (introduced in PR #46373, merged 2026-07-13) is flaky due to a timing race in the test harness, not a product bug. The test creates a workload manager classifier with a has_path = '/Root/et_s3' predicate and a resource pool with concurrent_query_limit = 0, then expects a query against an external table to be rejected with PRECONDITION_FAILED.

The WaitClassifierVisible helper polls the metadata table until the classifier appears, then adds a fixed 500ms sleep to wait for the workload service snapshot to refresh. Under CI load, 500ms may be insufficient — if the workload service snapshot hasn't refreshed by the time the test query executes, the classifier won't be active, and the query may return a different status instead of the expected PRECONDITION_FAILED.

Additionally, the has_path predicate requires a compiled query (it returns TPendingCompilation during pre-classification and resolves during post-classification). The test uses an ObjectStorage external data source pointing to https://storage.yandexcloud.net/tpc/. If query compilation encounters S3 connectivity issues or timeouts in CI, the query may fail with a different error status, bypassing the classifier match entirely.

This differs from sibling tests (DirectTopicMatches, CdcStreamMatches) which use local PQ/mock connectors and are therefore deterministic.

Flakiness window: The test first appeared on 2026-07-13 and showed flakiness within 24 hours (2026-07-14), with a 96.6% success rate (86 passed, 3 failed out of 89 runs). This is characteristic of a latent timing issue rather than a regression from a later change.

Suspect PR(s), ordered by causality (most likely first):

1. wm: it for has_path classifier over eds and topic #46373 — Introducing PR (wm: it for has_path classifier over eds and topic). The flakiness is inherent to the test design: insufficient wait for workload service snapshot refresh and dependency on external S3 connectivity during compilation. Author: buhtr (Slusarenko Igor).
2. wm: use string for ha_app_name, use glob for has_full_scan, has_path #46499 — (wm: make classifier predicates match the RFC). Changed has_path from regex to glob syntax on 2026-07-15. While the test uses a literal path /Root/et_s3 (no regex metacharacters), this change could theoretically affect matching behavior. However, the test passed on 2026-07-15 after this merge, making it unlikely.

Suggested Fix

1. Replace the fixed 500ms sleep with a polling loop: Instead of sleeping after WaitClassifierVisible, poll the workload service state directly or retry the ExecQuery assertion with backoff until the classifier is confirmed active. This is the approach used in other stabilized tests in this directory (see PR kqp/ut/federated_query/datastreams: stabilize tests #42552).

2. Mock the external data source: Replace the real ObjectStorage EDS pointing to https://storage.yandexcloud.net/tpc/ with a local S3 mock (an S3 recipe is already configured in ya.make). This would eliminate non-determinism from S3 connectivity during query compilation while preserving the has_path classification logic being tested.

3. Increase the snapshot refresh wait (quick mitigation): If a proper fix is deferred, increase the sleep in WaitClassifierVisible from 500ms to 2-3 seconds. This is a band-aid but would significantly reduce flake rate under CI load.

Summary

HasPathDatastreams.ExternalTableMatches is a newly introduced test (PR #46373, 2026-07-13) that is flaky due to a timing race in the workload manager classifier visibility wait (500ms fixed sleep insufficient under CI load) and a dependency on external S3 connectivity for query compilation. The fix should replace the fixed sleep with a polling loop and/or mock the ObjectStorage data source. Severity is LOW — 96.6% pass rate, no product bug, only the test harness needs adjustment. Responsible team: area/fq (@ydb-platform/fq), per TESTOWNERS for ydb/core/kqp/ut/federated_query.

[robot-vibe-db]

Full analysis log

Analysis performed by claude, claude-opus-4-6.